Nature Reviews Electrical Engineering (2026), DOI: 10.1038/s44287-026-00337-z

Transferable Predictive Maintenance For Battery Energy Storage Systems

Shengyu Tao¹, Torsten Wik¹, Changfu Zou¹, Chen Liang², Huadong Mo², Shiva Sander Tavallaey³, Niklas Thulin⁴, Enrico Zio ^{5, 6}, and Michael Pecht<sup>7

1</sup>Department of Electrical Engineering, Chalmers University of Technology, Göteborg, Sweden
²School of Systems and Computing, University of New South Wales, Canberra, Australian Capital Territory, Australia
³ABB Corporate Research, Västerås, Sweden
⁴Volvo Energy AB, Göteborg, Sweden
⁵Energy Department, Politecnico di Milano, Milan, Italy
⁶Mines Paris, Paris Sciences et Lettres University (PSL), Paris, France
⁷CALCE, University of Maryland, College Park, MD, USA

For more information about this article and related research, please contact Prof. Michael Pecht.

Abstract:

Battery energy storage systems are becoming multi-role infrastructures whose functions extend beyond energy storage and keep evolving over long deployment horizons. This evolution drives persistent operational drift in application requirements, duty cycles, environmental exposure and monitoring pipelines, whereas predictive maintenance still largely assumes fixed roles and stationary conditions. Drift therefore undermines the reliability of predictive maintenance, compromising system safety, asset availability and operational resilience. This Perspective proposes a system-level framework for transferable predictive maintenance under structured operational drift. We classify the sources of operational drift across evolving asset roles and organize their effects into layered functional mismatches spanning data, model and temporal layers. Each mismatch layer is linked to targeted transfer strategies, which we assess against negative transfer, catastrophic forgetting, calibrated uncertainty and sample efficiency. We position transferable predictive maintenance as a life-cycle-spanning capability connecting manufacturing, first-life operation and second-life deployment and outline the data infrastructures, foundation models, open benchmarks and regulatory recommendations needed to sustain it.

This article is available online here and to CALCE Consortium Members for personal review.